Background: Knowledge of the variability of a measurement method is essential for its clinical application. We investigated the variability of shunt flow measurements, since this is a relatively neglected area in the literature. In particular, no direct comparison of between-session and within-session variability was available until now.
Methods: During two consecutive dialysis sessions, shunt flow was measured three times with the ultrasound dilution method in 24 chronic haemodialysis patients with various types of shunts. Needle orientation and blood pressure at the time of flow measurement were recorded. In these patients, shunt flow was also measured three times by duplex ultrasound before the first dialysis session.
Results: The within-session variation coefficient (VC) of shunt flow measured with ultrasound dilution was 7.7%, whereas the between-session VC was 14.2% (n.s.). The within-session VC of Doppler shunt flow was 11.6% which was not significantly different from the corresponding figure of ultrasound dilution. Analysis of subgroups showed that changes in needle orientation caused large differences between sessions in radiocephalic fistulas but not in brachiocephalic fistulas: in the radiocephalic fistulas with the same needle orientation, VC was 6.7%, but with different needle orientation it was 23.5% (P = 0.02); the corresponding figures for brachiocephalic fistulas were 14.6% (same direction) and 11.4% (different direction, n.s.).
Conclusion: Reproducibility of shunt flow measurements between dialysis sessions in radiocephalic fistulas is critically dependent on similar needle orientation. With similar needle position and correction for blood pressure differences, flow changes of more than 20-25% are likely to reflect true flow changes. The variability of duplex flow measurements is at least as large as that of the ultrasound dilution method.